A semi-supervised anomaly detection strategy for drunk driving detection: a feasibility study DOI Creative Commons
Fouzi Harrou, K. Ramakrishna Kini, Muddu Madakyaru

et al.

Frontiers in Sensors, Journal Year: 2024, Volume and Issue: 5

Published: June 20, 2024

Drunk driving poses a significant threat to road safety, necessitating effective detection methods enhance preventive measures and ensure the well-being of users. Recognizing critical importance identifying drunk incidents for public this paper introduces an semi-supervised anomaly strategy. The proposed strategy integrates three key elements: Independent Component Analysis (ICA), Kantorovitch distance (KD), double Exponentially Weighted Moving Average (DEWMA). ICA is used handle non-gaussian multivariate data, while KD measure dissimilarity between normal abnormal events based on features. DEWMA applied charting statistics detect changes in data uses nonparametric threshold improve sensitivity. primary advantage approach its ability perform without requiring labeled data. study also XGBoost later calculation SHAP (SHapley Additive exPlanations) values identify most important variables detecting behavior. was evaluated using publicly available from gas temperature sensors, as well digital cameras. results showed that achieved F1-score 98% driver’s status, outperforming conventional PCA-based ICA-based methods.

Language: Английский

Enhanced prediction of ventilator-associated pneumonia in patients with traumatic brain injury using advanced machine learning techniques DOI Creative Commons

Negin Ashrafi,

Armin Abdollahi,

Kamiar Alaei

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 2, 2025

Language: Английский

Citations

1
Hierarchical manifold sample envelope transformation model for ensemble classification DOI
Jie Ma, Hong Chen, Yongming Li

et al.

Computers & Electrical Engineering, Journal Year: 2025, Volume and Issue: 123, P. 110252 - 110252

Published: April 1, 2025

Language: Английский

Citations

0
Interpersonal Sensitivity Prediction Based on Multi-strategy Artemisinin Optimization with Fuzzy K-Nearest Neighbor DOI
Yingjie Tian, Xiao Pan,

Xinsen Zhou

et al.

Journal of Bionic Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: April 13, 2025

Language: Английский

Citations

0
A semi-supervised anomaly detection strategy for drunk driving detection: a feasibility study DOI Creative Commons
Fouzi Harrou, K. Ramakrishna Kini, Muddu Madakyaru

et al.

Frontiers in Sensors, Journal Year: 2024, Volume and Issue: 5

Published: June 20, 2024

Drunk driving poses a significant threat to road safety, necessitating effective detection methods enhance preventive measures and ensure the well-being of users. Recognizing critical importance identifying drunk incidents for public this paper introduces an semi-supervised anomaly strategy. The proposed strategy integrates three key elements: Independent Component Analysis (ICA), Kantorovitch distance (KD), double Exponentially Weighted Moving Average (DEWMA). ICA is used handle non-gaussian multivariate data, while KD measure dissimilarity between normal abnormal events based on features. DEWMA applied charting statistics detect changes in data uses nonparametric threshold improve sensitivity. primary advantage approach its ability perform without requiring labeled data. study also XGBoost later calculation SHAP (SHapley Additive exPlanations) values identify most important variables detecting behavior. was evaluated using publicly available from gas temperature sensors, as well digital cameras. results showed that achieved F1-score 98% driver’s status, outperforming conventional PCA-based ICA-based methods.

Language: Английский

Citations

0